The Measurement Trap

Is it time to rethink how and what we measure, especially in the credit union world where “high performance” has a different meaning than in for-profit financial institutions?


“You must be able to measure progress to achieve it” is a management axiom. There's even a school of thought that argues, “If you can’t measure it, it doesn’t matter.”

Most thinkers have moved past that limited view, but the notion that performance requires the discipline of cardinality is embedded deep in our management DNA. This is why a recent blog post from a Ph.D. candidate in finance at the University of South Florida caught my eye.

In the post, Leo Chen quotes Goodhart’s Law, “When a measure becomes a target, it ceases to be a good measure.”

It’s easy to see why this never achieved the status of Moore’s Law or Occam’s Razor.  It’s nuts, right? 

To be fair, Charles Goodhart, who was an economist at the Bank of England and a professor at the London School of Economics, was thinking in the macro world. As first formulated 40 years ago, his law refers to large measures — interest rates, unemployment, homelessness, and the like — and their relationships with government policy. Chen, who was writing for investors, applies Goodhart as a caution against using widely quoted market indicators as baselines against which to judge investment performance.

Could the use of traditional management measures actually be so distortive as to be counterproductive? 

Still, with spring strategy season just around the corner, I find myself wondering about Goodhart’s more general application. Could the use of traditional management measures actually be so distortive as to be counterproductive? What would the implications of this be for credit unions? Credit unions work for members, not third-party stockholders, but they still depend on a vast array of traditional measures to set targets and evaluate results. Could this be a problem?

In the for-profit world, sales is probably the most obvious example of targets becoming bad measures. Sales managers live the nightmare of getting from their sales forces exactly what they incentivize. Not what they want. Not what they intend. What they actually measure and reward.

Want five new accounts this month? Done! Thanks for the bonus.

You want them to be profitable? Better include that in the incentive plan.

Likewise, there is an entire executive compensation literature based on the same point: If you're paying your CEO with stock options, you're providing an incentive to maximize the share price on the strike date, regardless of the long-term impact on the company of doing so.

Common Sense?

Taken in this light, Goodhart’s Law doesn’t seem so nuts. Provocative perhaps, deeply disturbing, maybe even subversive, but not crazy. To the contrary, it forces a rethink of foundational management assumptions. Or maybe it’s just common sense: Using a good measure of performance as a performance target creates an incentive to manipulate the measurement, degrading its value.

Your board wants more productivity? Its members are probably looking for more top-line income without a concomitant rise in expenses, but the odds are that they'll express the goal and measure progress in terms of the efficiency ratio. The easy approach is to cut expenses — defer some maintenance, cut back on training, slow down hiring, etc. We all understand those are the wrong things to do, but they do produce the sought-after results. 

Simple popular benchmarks are problematic because the simplicity of tweaking them to present undesirable outcomes in attractive terms creates a moral hazard.

Unfortunately, most popular measures of performance are ratios, and institutions can improve any ratio by changing either one of the inputs. This makes Goodhart’s Law as applicable to micro examples of performance measurement as it is at the macro level. Simple popular benchmarks are problematic as management tools because the simplicity of tweaking them to present undesirable outcomes in attractive terms creates a moral hazard if managers employ the benchmarks as performance targets. 

Of course, we cannot stop measuring things just because basic measures are easily manipulated, but perhaps we need to rethink how and what we measure, especially in the credit union world where high performance has a different meaning than in for-profit financial institutions. 

Return Of The Member

The challenge, then, is a benchmark that measures virtue but cannot be easily manipulated to make venality look virtuous. One such measure in the credit union world is Return of the Member. Users of Callahan & Associates Peer-to-Peer data analytics might recognize the initials ROM, and its implications are clear. Its components, by design, are not so clear. As a result, it presents an exception to Goodhart’s Law, not lending itself to easy manipulation that would degrade its value as an honest indicator of performance.

ROM is simple in conception but highly complex in construction. It is designed to capture the cross-section of performance elements that maximize the value of a financial cooperative for its member-owners. ROM has three core components: Return to Savers, Return to Borrowers, and Member Service Usage. Each comprises a variety of measurements and performance ratios weighted to give a compelling picture of how well a given credit union serves its members and performs in these vital tasks relative to its peers and competitors.

Want More ROM?

Click here to learn more about how Callahan calculates Return of the Member as well as what makes a ROM leader.


WHat is rom   Rom leaders

ROM considers service quality, but not at the expense of cost management, which is also measured in several ways. Likewise, ROM accounts for lending standards, but also risk management.

It’s not that credit unions cannot manipulate their ROM scores. They can. But the balance and detail of the ROM metric captures and accounts for the tradeoffs that come with manipulation and other management choices. Raise your savings rates to attract more deposits and your return to savers will increase, but that money came from somewhere, and the change in priorities will be reflected elsewhere in the measure. 

At its core, ROM is a subjective collection of objective measures. It represents philosophical judgments about how and why we are in this business while providing concrete means of determining progress and comparative performance. Any compromise between value judgments and cardinal scales will embody tradeoffs, but a balanced compromise will account for and leverage countervailing inputs, effectively defying Goodhart’s Law. By my yardstick, that’s pretty good for a management tool.

Chris Howard is vice president of research at Callahan & Associates. An experienced board and management consultant with more than two decades in financial services, he writes and speaks on strategy, planning, innovation, management, and big data. Reach him directly at


Feb. 24, 2015



No comments have been posted yet. Be the first one.